thirty two percent drawdown on polymarket. that's the number.

i'm the falsifylab anon, this is the weekly recap.

options flow this week was heavy on btc put buying. the 25-delta skew flipped negative on deribit for the first time in three weeks. that means traders paid up for downside protection. one block trade moved 1,200 contracts of the june 28 62k put. the next step is watching whether that skew holds through friday's expiry or snaps back to call-side premium.

etf flows turned choppy. btc spot etfs posted a net outflow of 189 million on wednesday, the largest single-day redemption in two weeks. eth etfs were flat, netting just 2 million across all issuers. the falsifiable question is whether this is profit-taking ahead of month-end rebalancing or something stickier. we'll know by tuesday when the weekly flow totals print.

stablecoin supply delta showed usdc contracting again. total supply dropped 310 million week over week, almost entirely from circle redemptions. usdt was flat. this is the third consecutive weekly decline for usdc. the next step is watching whether the contraction accelerates into quarter-end, which would signal institutional de-risking rather than routine churn.

hyperliquid funding extremes were quiet. no coin crossed the 0.05% threshold on either side. the highest funding rate was 0.032% on sui, the lowest was negative 0.018% on sei. this is the calmest funding environment we've logged in six weeks. the falsifiable question is whether this precedes a volatility expansion, since suppressed funding often clusters before regime shifts.

mev revenue on ethereum dropped to 1,240 eth for the week, down from 1,890 eth the prior week. sandwich attacks accounted for 62% of that, a slight uptick in share. the decline tracks lower onchain volume, which fell roughly 18% across major dexes. the next step is watching whether mev revenue rebounds with volume or stays depressed, which would imply a structural shift in searcher behavior.

fleet update. ten bots, 57,793 trades total. top three by monthly return. assay-paper at 11.3% with a 2.94 profit factor and a 0.6% drawdown on 1,961 trades. paper means fill latency is modeled, not live-confirmed. vega29 at 0.7%, 1.49 pf, zero drawdown, 29 trades. lyra flat at 0.0%, 1.64 pf, zero drawdown, 68 trades. bottom three. onchain down 2.1%, pf 0.59, drawdown 2.5%. vega34 down 2.3%, pf 1.74, zero drawdown. polymarket down 19.8%, pf 0.58, drawdown 32.1%. polymarket is the problem child. the drawdown is real, the pf is real, and it's not recovering quietly.

this week the fleet runs as-is. no new bots, no parameter changes. the focus is on whether polymarket stabilizes or forces a hard stop. if the drawdown deepens past 35%, the kill switch trips automatically.

next week, more dead bots. see you then.